The general purpose of this research is to establish electrophysiologic correlates of psychiatric illness. Such correlates may then serve to focus animal research on underlying mechanisms. EEG and evoked response data are obtained in different clinical groups and analyzed in various ways with computer assistance. Patients are followed through treatment with serial determinations to assess the effects of drugs. Evoked response experimental procedures currently employed include: (1) Pseudorandomized presentation of electrical pulses to right and left wrists, full field visual checkerboard flash, binaural click, with monopolar recording from 15 leads; on-line averaging is in two modes, serial and intermittent, to assess short- and longer term variability. (2) Recovery functions, using one interval (10 msec) and varying intensity of single conditioning stimuli or trains. EEGs are subjected to power spectrum and amplitude and frequency time series analyses. Evoked responses are automatically analyzed for amplitude, latency, and wave shape stability, and may be digitally filtered to determine which frequency components relate to independent criteria. Reduced data may be treated statistically with factor analysis or canonical correlation procedures. Some recent results of interest are: (a) Activity occurring more than 100 msec after stimulus is generally attenuated in schizophrenic patients. (b) Normal negative peaks at 130 msec in the somatosensory potential are absent in the majority of schizophrenics; in the auditory potential normal negativity at 100 msec occurs earlier in schizophrenics and is attenuated. (c) A somatosensory negativity peaking at 60 msec is augmented in postcentral recordings in chronic paranoid and chronic undifferentiated schizophrenics, but not in other subtypes.